Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Which program can solve position from Nisipeanu-Khalifman?

Author: Jeremiah Penery

Date: 06:43:22 10/12/99

Go up one level in this thread


On October 12, 1999 at 09:26:43, Bernhard Bauer wrote:

>On October 12, 1999 at 09:10:40, Jeremiah Penery wrote:
>
>>On October 12, 1999 at 08:47:57, Bernhard Bauer wrote:
>>
>>>Hallo,
>>>
>>>       +---+---+---+---+---+---+---+---+
>>>    8  |   |   |   |   |   |   | *K|   |
>>>       +---+---+---+---+---+---+---+---+
>>>    7  |   | R |   |   |   |   |   |   |
>>>       +---+---+---+---+---+---+---+---+
>>>    6  |   |   |   |   |   |   | P | *P|
>>>       +---+---+---+---+---+---+---+---+
>>>    5  |   |   |   | *R| *B| *P|   | P |
>>>       +---+---+---+---+---+---+---+---+
>>>    4  |   | P |   |   |   | B |   |   |
>>>       +---+---+---+---+---+---+---+---+
>>>    3  |   |   |   |   |   |   | K |   |
>>>       +---+---+---+---+---+---+---+---+
>>>    2  |   |   |   |   |   |   |   |   |
>>>       +---+---+---+---+---+---+---+---+
>>>    1  |   |   |   |   |   |   |   |   |
>>>       +---+---+---+---+---+---+---+---+
>>>         a   b   c   d   e   f   g   h
>>>FEN: 6k/1R/6Pp/3rbp1P/1P3B/6K/8/8 w
>>>
>>>this is a position from Nisipeanu vs. Khalifman, Las Vegas, tie-break, game 2.
>>>
>>>Question: which program can solve this position?
>>>IMHO no program can do, but who knows, so I ask.
>>>G. Kasparov calls it a mutual Zugzvang position and in the variant he gives the
>>>white rook is hold in prison.
>>>The best move G. K. gives is Kf3 !!
>>
>>I assume this move wins for white?  What is the variation?
>>
>>Just getting Kf3, my modified Crafty 16.18 does this in 1:46 (less than 2
>>minutes) at ply 13.  This is the variation at ply 14 and 2 1/2 minutes:
>>
>>14     2:34   6.12   1. Kf3 Bxf4 2. Kxf4 Rd1 3. Kxf5 Rf1+
>>                     4. Ke6 Re1+ 5. Kf6 Rf1+ 6. Ke7 Re1+
>>                     7. Kd6 Rd1+ 8. Ke5 Re1+ 9. Kd5 Rd1+
>>                     10. Kc5 Rc1+ 11. Kd4 Rd1+ 12. Ke4 Kf8
>>
>>I'll let it run all of today and see what happens.
>>
>>Jeremiah
>
>G. K. writes:
>"Only 40.Bxe5?? blew up everything - position after 40...Rxe5 41.Kf4 Rd5
>was a mutual Zugzvang! 42.Rh7 doesn't work - 42...Rd4+ 43.Kxf5 Rxb4 44.Rxh6 >Rb7!

I wonder how long it would take for the score to drop.  When mine changed to
Kf3, the score for Bxe5 was still almost +6.  The score change was +.39, but
then dropped immediately so that the score change from Bxe5 to Kf3 was only +.26
pawns.

>That's why clever 40.Kf3! would have finished the job - 40...Bxf4 41.Kxf4 and
>it's black to move! E.g. 41...Kf8 42.Rh7 Rd4+ 43.Ke5 Rxb4 44.Rxh6 and there is
>no more Rb7 cutting white rook on h6."

I'm not sure what's going on with my PV...The white king makes an interesting
march, and the black checks can go on for quite some time.  But eventually the
white king will hide behind the b-pawn, then use the rook for further hiding
while he runs the pawn, I think.  It's an interesting problem, and highlights
some of the computer difficulties in the endgame.

>My crafty (v16.19) gives:
>                                    10. Rh7+ Kf6
>               15->   7:08   5.15   1. Bxe5 Rxe5 2. Kf4 Rd5 3. Rh7 Rd4+
>                                    4. Kxf5 Rxb4 5. Rxh6 Rb7 6. Kf6 Ra7
>                                    7. Ke6 Ra6+ 8. Kd5 Ra5+ 9. Ke4 Kg7
>                                    10. Rh7+ Kf6


Interesting...It sees the correct line leading toward the mutual zugzwang, but
doesn't choose Kf3...I wonder how many more ply it would take.

Mine did think for a long time about 40. Re7, but didn't select it.  It's still
thinking about ply 15 after 30 minutes.

By the way, what sort of computer are you running on?  It seems to be quite
fast. :)

Jeremiah



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.